Antibody Therapeutics in Cancer

See allHide authors and affiliations

Science  13 Sep 2013:
Vol. 341, Issue 6151, pp. 1192-1198
DOI: 10.1126/science.1241145


In a relatively short period of time, monoclonal antibodies have entered the mainstream of cancer therapy. Their first use was as antagonists of oncogenic receptor tyrosine kinases, but today monoclonal antibodies have emerged as long-sought vehicles for the targeted delivery of potent chemotherapeutic agents and as powerful tools to manipulate anticancer immune responses. With ever more promising results from the clinic, the future will likely see continued growth in the discovery and development of therapeutic antibodies and their derivatives.

Since the discovery in 1948 that cytotoxic folate antimetabolites could be used in the treatment of childhood leukemia, our basic approach to cancer therapy has remained fundamentally the same: surgery followed by sublethal administration of various cytotoxic compounds or radiation. However, over the past 16 years the situation has started to change substantially with the advent of “targeted” cancer therapies: the use of drugs developed to inhibit oncogenic proteins or survival factors selectively expressed by tumors. Although targeted agents are often thought of as low molecular weight inhibitors that can be given orally and gain access to cytoplasmic targets by diffusion across the plasma membrane, the first and perhaps most effective targeted therapies involve the use of biotherapeutics, recombinant proteins (usually antibodies) that modulate targets expressed at the cancer cell surface. This is quite remarkable because, not so long before the introduction of the first effective therapeutic antibodies in cancer (the CD20 antibody rituximab and the HER2 antagonist trastuzumab), there were severe doubts that antibodies could ever be safely used in humans. One major impediment was that monoclonal antibodies were typically produced in mice and therefore would be expected to be immunogenic in humans. The discovery that one could humanize murine antibodies by grafting complementarity-determining regions (CDRs) from the desired mouse antibody onto a recombinant human immunoglobulin backbone changed everything (1, 2).

Today, a variety of sophisticated strategies exist to engineer humanized antibodies and to produce human antibodies by using rodents partially reconstituted with human immunoglobulin genes, human immune cells, or combinatorial phage or yeast display libraries that can yield high-affinity antibodies without requiring animal immunization at all (3). These approaches have led to an explosion of therapeutic antibodies both in cancer and in other diseases (4). Thus far, the only Food and Drug Administration (FDA)/European Medicines Agency–approved agents for oncology are conventional immunoglobulin G (IgG) molecules or their conjugates (Fig. 1), but engineered antibodies and novel antibody-like variants are increasingly making their way into clinical trials.

Fig. 1 Space-filling model of IgG1.

Light chain, magenta; heavy chain, dark blue; carbohydrate, yellow. [Credit: C. Eigenbrot, Genentech, Incorporated]

In this review we will discuss the general classes of antibody therapeutics that are commonly in use, provide a detailed analysis of two specific examples of oncogene-targeted antibody drugs, and consider some of the most promising new strategies, namely the use of antibodies for targeted delivery of conventional chemotherapeutics and to enlist the power of the immune system.

Classes of Antibody Therapeutics in Cancer

There are currently 13 antibodies approved by the FDA for various oncology indications (Table 1), and many more are currently being evaluated in clinical trials. Antibodies for the treatment of hematologic cancers, such as non-Hodgkin’s lymphoma or chronic lymphocytic leukemia, and antibodies to B cell–associated targets (CD20, CD52) act, at least in part, because of their ability to induce apoptosis after binding to B cell tumors (5). The first of the approved antibodies directed against targets expressed on solid tumors were selected in cell culture as antagonists of the oncogenic receptor tyrosine kinases epidermal growth factor receptor (EGFR) and HER2 (6).

Table 1 Antibody therapies and their indications.
View this table:

One antibody, bevacizumab, binds and sequesters vascular endothelial growth factor–A (VEGF-A), a key factor that is required for the growth of blood vessels and exerts its antitumor effect by functionally altering or slowing the formation of the tumor vasculature (7). Although this is at present the only approved antibody in oncology that targets a secreted protein, this class of therapeutic is common in the case of chronic inflammatory diseases, where multiple approved agents target proinflammatory cytokines such as tumor necrosis factor–α (4).

Two antibodies are now approved that are covalently modified with cytotoxic microtubule antagonists. These antibody-drug conjugates (ADCs), one of which is a modified version of trastuzumab for HER2-positive breast cancer patients (8), enable a form of targeted chemotherapy. These recent successes have led to a plethora of ADCs in clinical development (see below).

Last, the recent approval of ipilimumab has introduced yet another therapeutic strategy for anticancer antibodies, namely active immunotherapy (9). Ipilimumab blocks the activity of the negative regulator of T lymphocytes, CTLA4, resulting in the activation of a patient’s immune response to cancer. This exciting approach has already demonstrated long-term, durable benefit.

Most of the approved antibodies in oncology are of the human IgG1 subclass, the subclass that is the most effective at engaging Fcγ receptors (FcγRs) on natural killer (NK) cells, macrophages, and neutrophils (Table 1). Antibody engagement of these receptors leads to the killing of antibody-bound target cells by antibody-dependent cellular cytotoxicity (ADCC) or antibody-dependent phagocytosis (ADP) (1012). This is an important consideration given that ADCC or ADP likely contributes to the efficacy of these antibodies, above and beyond any direct effect on modulating signal transduction. FcγR-mediated cross-linking of surface-bound antibodies may be important in some settings, for example, to induce apoptosis (13).

The next generation of antibody therapeutics will likely extend beyond the IgG1 isotype and/or include modifications to the Fc domain, the region that binds to FcRs. Where it may be beneficial to reduce ADCC, the IgG4 subclass has been used, although here the hinge region of the antibody must be engineered to prevent premature dissociation and reassembly of heavy chain–light chain half molecules with endogenous IgG4s in the plasma (14). Complete elimination of ADCC can be achieved by modifying specific residues in the Fc domain that bind to FcγR or by producing recombinant antibodies that lack the N-glycosylation of heavy-chain Fc regions, required for recognition by FcγRs (15). Conversely, alterations in glycosylation have been identified that enhance FcγR binding (effector domain–enhanced antibodies), thereby increasing the potential for ADCC or FcγR cross-linking–dependent signaling. Still other Fc domain mutations are being tested that either enhance or reduce binding the neonatal FcR for IgG (FcRn), the receptor responsible for returning IgGs to the circulation after their internalization by endothelial cells. Thus, enhanced FcRn binding will serve to increase the half-life of injected antibodies, whereas reduced FcRn binding will decrease half-life (16). These may prove to be important considerations in controlling the pharmacokinetic exposure levels of a given antibody, with a potential toxicity possibly mitigated by faster clearance.

New formats are also being developed, such as two types of bispecific antibodies. In the first, single heavy chain–light chain pairs derived from antibodies with distinct specificities can be covalently or noncovalently linked to yield a single IgG molecule bearing arms targeted to two different antigens. To overcome the issue of monovalent binding that comes with this approach, phage display has been used to make single CDRs that exhibit dual specificities. Such dual-action Fabs are otherwise normal antibodies that retain high affinity and bivalent affinity for two antigens simultaneously. One such antibody (MEHD7945A) that detects EGFR and the HER3 co-receptor simultaneously is currently in phase II clinical trials (17). In other cases, one-armed antibodies (i.e., an intact IgG missing one Fab domain, onartuzumab) have been deployed where it was desirable to avoid the cross-linking of a potentially activating receptor (c-Met) (18).

Even less traditional formats are being evaluated, involving attenuated single-chain antibodies that have either single or dual specificities (19). These antibodies have the limitation that smaller antibody constructs have shorter in vivo half-lives after injection, which necessitates frequent or continuous injections. Larger, multiple-armed antibodies or antibodies genetically fused to other mediators (e.g., cytokines, hormones, and toxins) as well as a variety of synthetic “minibodies” and antibody mimetics are also on the horizon and, if successful, could increase the potential range of therapeutic strategies. However, one caution is that the farther one strays from the conventional, the greater the chances are for unanticipated behavior, adverse effects, or inherent immunogenicity.

EGFR Therapeutics

The EGFR family (ErbB/HER) has proved to be a fruitful area for the development of cancer therapeutics. To date, eight agents are approved for the treatment of various solid tumors, and at least a dozen other compounds are currently in clinical development.

EGFR therapeutics are divided into two classes, both of which inhibit EGFR: monoclonal antibodies (mAbs), cetuximab and panitumumab, and low molecular weight tyrosine kinase inhibitors (TKIs), gefitinib and erlotinib. mAbs directed against EGFR were first developed in advanced colorectal cancer. The pivotal trial for cetuximab demonstrated tumor shrinkage when used in combination with the topoisomerase I inhibitor irinotecan in patients whose tumors were progressing on irinotecan-based therapy (20). Discernable single-agent cetuximab activity was also observed. The murine parent mAb, 225, was reengineered as a human chimera, which allows for ADCC. In contrast, panitumumab was derived from a transgenic mouse capable of producing human antibodies and exhibits considerably higher binding affinity for EGFR than cetuximab. Panitumumab is a human IgG2, which is less effective in mediating ADCC than IgG1 (21). Both antibodies are restricted to use in patients whose tumors overexpress EGFR and are wild type for the proto-oncogene KRAS but have not yet been compared directly, making it impossible to judge the contribution of ADCC.

Although both EGFR mAbs and small-molecule inhibitors have been commercially available for more than a decade, there is remarkably little overlap in how the two classes of antagonists are used clinically. Combining mAbs and small molecules is ineffective, despite preclinical data suggesting increased efficacy (22). Similarly, antagonism of EGFR was originally predicted to synergize with cytotoxic chemotherapy. Combination therapy of irinotecan with cetuximab was key to cetuximab FDA approval. Yet erlotinib or gefitinib failed to show clinical benefit in combination with several chemotherapies (23, 24). Indeed, there is a suggestion that patient outcome may be worse when treated with the combination (25). Many factors contribute when preclinical observations fail to predict clinical responses, including differences in pharmacokinetics and, most importantly, differences in tolerability. Although both mAb and chemical inhibitors of EGFR cause skin rash, erlotinib and gefitinib cause additional toxicities that further limit tolerability when combined with conventional chemotherapies. Of interest in this regard is the recent clinical demonstration (and FDA approval) that erlotinib is more effective against tumors expressing an activated mutant allele of EGFR (24). Although both wild-type and mutant kinases are inhibited, selectivity for the mutant kinase might be expected to increase therapeutic index by sparing, relatively, the activity of the wild-type kinase in normal cells and tissues.

HER2 Therapeutics

Three mAbs (trastuzumab, pertuzumab, and ado-trastuzumab emtansine) and a small-molecule TKI (lapatinib) are approved for the treatment of HER2-positive breast cancer (Table 1), yet interpretation of the clinical data is less complex than that observed with EGFR antagonists. Lapatinib, a low molecular weight HER2 and EGFR TKI, effectively blocks HER2 signaling and was approved on the basis of compelling clinical data generated in patients whose tumors had progressed on trastuzumab-based therapy (26). Thus, HER2 tyrosine kinase antagonism alone is an effective treatment for this form of breast cancer. Indeed, trastuzumab and pertuzumab, antibodies against HER2, were selected for their ability to block proliferation (27) and signaling activity (28). Although these antibodies were also humanized to IgG1, suggesting that ADCC might add to their clinical activity, human FcγR polymorphisms that enhance binding are not linked to improved responses in patients (29). Yet the relative importance of ADCC remains an active area of debate.

In contrast to EGFR antagonists, HER2-directed antibodies and small molecules are very effective in the treatment of HER2-positive breast cancer in combination with conventional chemotherapy. Because of favorable tolerability profiles, HER2 targeting can also be combined with standard cytotoxic chemotherapy. Indeed, pertuzumab is approved only in combination with trastuzumab and docetaxel (30). As is often the case in cancer therapy, the ability to combine multiple drugs produces better clinical outcomes; combinations are enabled by better safety profiles.

Because trastuzumab and pertuzumab bind to different regions of HER2 (Fig. 2), dual antibody therapy should allow for simultaneous antagonism of both activated forms of HER2. Support for this hypothesis was initially obtained in a phase II clinical trial that tested combination antibody therapy in patients whose tumors progressed on trastuzumab and cytotoxic chemotherapy. The findings in this trial demonstrated an objective response (tumor shrinkage) of 25% and a clinical benefit rate of 50% (31). A neoadjuvant (i.e., treatment before surgery) trial further confirmed that the antibody combination was also effective in patients who were naïve to therapy (32). Equipped with the knowledge that dual HER2 antibody blockade was active in early breast cancer and late-stage metastatic breast cancer, a pivotal trial validated the approach for the treatment of first-line metastatic breast cancer patients and led to the accelerated, full approval of pertuzumab in June 2012 (30). Indeed the magnitude of patient benefit, as measured by progression-free survival (3.9 versus 6.1 months) (33) and overall survival (34), is greater than what was observed for the initial approval of trastuzumab.

Fig. 2 Epitope binding of HER2 and EGFR therapeutic antibodies.

(A) Trastuzumab. (B) Pertuzumab. (C) Cetuximab or panitumumab. (D) T-DM1 (ado-trastuzumab emtansine). In the case of HER2 signaling, a functional and physical association with a second related receptor, HER3, is often important (60). HER2 kinase phosphorylates the HER3 cytoplasmic domain, activating it as a scaffold to promote the phosphatidylinositol 3-phosphate kinase cascade. Conversely, activation of HER3 by binding ligands such as heregulin can allosterically activate HER2 kinase. Pertuzumab’s likely mechanism of action is to block heterodimerization of HER2 and HER3. Asterisks indicate a cytotoxic agent, derivative of maytansine 1 (DM1).

In summary, the advent of mAbs to oncogenic receptor tyrosine kinases of the EGFR family has represented a major advance in cancer therapy. Yet, there have also been attempts that have not yet led to success in the clinic, using mAbs to target other signaling receptors. Notable among these are antibodies to insulin-like growth factor 1 receptor, despite this receptor having a mechanistic link to many types of cancer. On the other hand, antibodies to c-Met continue in clinical development. The number of mAbs tested in clinical trials is small, which makes it difficult to draw any clear lessons yet. However, it should be noted that effective antibodies against solid tumors tend to recognize targets that are either subject to mutation (EGFR, c-Met) or overexpression (HER2) in cancer, even if the antibody itself is inactive against the mutant allele. Conceivably, mAb therapy against signaling receptors may prove most effective in cases where targets can be shown genetically to be important drivers of cancer or clear drivers of resistance to other forms of therapy.

Antibody-Drug Conjugates (ADCs)

The concept that antitumor antibodies could be used as vehicles for the selective delivery of cytotoxic agents to tumors has been around nearly as long as mAbs. However, until very recently, the idea has eluded successful implementation, probably for three reasons: (i) the use of antibodies against targets that were not sufficiently restricted to tumor cells, (ii) the use of drugs with insufficient potency or (in the case of bacterial or plant toxins) that were immunogenic, and (iii) the linker chemistry used to attach drugs to antibodies was not optimized. The latter consideration is one that required an enormous empirical effort to solve. A linker that was too labile allowed too much of the toxic drug to dissociate from the antibody in the blood. This then exposed normal tissues to the freed drug and, as a result, reduced the doses of the ADC that could be used. Stable linkers require complete proteolytic digestion of the ADC, releasing the derivatized amino acid, linker, and cytotoxic drug as the active metabolite (35). Similarly, the choice of drug required a great deal of trial and effort, with current successes being seen by using microtubule antagonists.

Gemtuzumab ozogamicin was the first ADC to reach the market in 2000. It is directed against CD33, a surface glycoprotein characteristic of acute myeloid leukemia (36). The conjugate is composed of a humanized IgG4, a pH labile linker, and calicheamicin, a very potent DNA minor groove binder. Unfortunately, the drug was withdrawn from the market in 2010 because of the failure of subsequent confirmatory studies to reproduce the initial clinical benefit profile (37, 38).

More recently, brentuximab vedotin, a mAb against CD30 that was covalently modified with the microtubule antagonist monomethyl auristatin E (MMAE), was approved for use in advanced Hodgkin’s lymphoma given its efficacy and good safety profile (39). The linker used to tether MMAE to the antibody contains a peptidic moiety that was designed as a substrate for cathepsin B; the antibody was delivered to cathepsin B–containing endosomes and lysosomes after internalization by the CD30-positive lymphoma cells (40).

Another ADC that has also recently gained approval is T-DM1, which is composed of trastuzumab (T) conjugated to a maytansine derivative (DM1) (41). In this instance, the linker is designed not to be cleaved by lysosomal proteases, suggesting that the drug is released only after limited digestion of the entire antibody. T-DM1 was found to prolong progression-free survival (9.6 versus 6.4 months) and overall survival (30.9 versus 25.1 months) in late-stage HER2-positive breast cancer patients who progressed on multiple other treatments (42).

The clinical successes of T-DM1 and brentuximab vedotin have stimulated a great deal of activity in developing ADCs to different solid and hematologic tumor targets using mAbs to a variety of antigens, most often conjugated to microtubule antagonists (MMAE and DM1 or their derivatives). In addition, new toxic payloads and linkers are under development in an effort to further increase therapeutic index and provide for combination therapy.

Novel engineered antibody platforms are also being investigated. For example, the typical cleavable MMAE linker drug (vc-MMAE) is coupled to antibodies after the reduction and then the derivatization of endogenous cysteine residues normally engaged in interchain disulfide bonds. Methods are being developed to engineer specific conjugation sites (e.g., unpaired cysteine residues) and defined sites on the heavy or light chains, leading to greater ease of conjugation and control over the number of drug molecules per antibody molecule (43). Other modifications are also being investigated in an effort to increase specificity of cancer cell binding (e.g., use of bispecific antibodies that recognize two antigens at once) or decrease the nonspecific uptake of the ADCs by normal cells that do not express the targets. Use of antibody-conjugated nanocarriers loaded with cytotoxic agents is also under investigation. Although this would increase the payload delivered after endocytosis, a potential limitation is the nonspecific loss of the cytotoxic in the plasma. We anticipate that the next several years will witness explosive growth and innovation in this area.

Antibodies and Cancer Immunotherapy

As was true for ADCs, the concept that a patient’s immune system could be activated to generate anticancer responses has been actively explored for >30 years. For most of this time, however, little progress was made in the clinical arena, largely because of our incomplete knowledge concerning the functional organization of the immune system and how tumors manipulate the immune response. Moreover, much of the field was long fascinated by anticancer “vaccines,” an approach limited by the immunosuppressive tumor microenvironments (44). Many cancer patients make anticancer T cell responses because tumors nearly always express mutant proteins or antigens to which the immune system is not tolerized and therefore are immunogenic. Yet these responses rarely manifest in clinical benefit.

An appreciation has now emerged that overcoming immunosuppression and enhancing the quality of the immune response may be key to developing effective cancer immunotherapies (3). This realization has generated enormous interest and excitement fueled by the early clinical successes of two classes of mAbs. The first of these is ipilimumab, which inhibits CTLA4, a negative regulator expressed on the surface of all T cells (Fig. 3A) (45). CTLA4 normally binds to CD80/CD86 on the surface of antigen-presenting cells, providing a powerful brake or checkpoint that limits the activation and proliferation of T cells after recognizing their cognate antigen. Ipilimumab blocks this interaction, thereby releasing the checkpoint and amplifying T cell responses. CTLA4 blockade also induces the death of regulatory T cells, which contribute to the immunosuppressive tumor microenvironment, by a mechanism that is still being characterized (possibly ADCC in the tumor bed; ipilimumab is an IgG1) (46, 47). In the clinic, ipilimumab was found to provide robust and long-term survival benefit or cures to patients with late-stage metastatic melanoma who otherwise had no other treatment options (9). Although the percentage of patients receiving benefit was small (3-year survival of 23.5% versus ~10% in a control arm) and cannot yet be diagnostically defined, the results are impressive and led to ipilimumab being approved by the FDA in 2011. Subsequently, ipilimumab has been reported to exhibit varying levels of clinical activity in other indications, but its use is complicated by the fact that it also induces occasionally serious immune-related adverse events (e.g., autoimmune inflammation). In any event, these studies established critical proof of concept for checkpoint inhibitors specifically and for other immunomodulators in general.

Fig. 3 The use of antibodies in active and passive immunotherapy of cancer.

(A) Checkpoint blockade by anti-CTLA4. CD80 and CD86 are ligands expressed on the surface of activated dendritic cells during the presentation of MHC [human lymphocyte antigen (HLA)]–peptide complexes to T cell receptors. CD80/86 binds to the costimulatory molecule CD28 to help activate T cell proliferation and then to the checkpoint inhibitor CTLA4 to attenuate T cell proliferation. The antibody ipilimumab blocks the interaction of CTLA4 with its ligands, thereby releasing the checkpoint inhibitor and favoring T cell proliferation. (B) Checkpoint blockade and inhibiting immune suppression by anti-PD1 or anti–PD-L1. (Left) T cell influx into tumors results in the release of IFN-γ, which up-regulates PD-L1 expression by tumor cells. PD-L1 binds to PD-1, which is expressed by activated T cells, generating a negative signal that causes T cell exhaustion (inhibiting the ability of T cells to recognize and kill their targets). (Right) During antigen presentation by dendritic cells, PD1 can also act as a checkpoint inhibitor where a negative signal can be sent by its binding to either PD-L1 or the closely related (and dendritic cell–specific) negative regulatory ligand PD-L2. Generally, inhibition of both PD-L1 and PD-L2 (e.g., by anti–PD-1) is required to block negative regulation by dendritic cells, whereas only PD-L1 inhibition (by anti-PD-1 or anti–PD-L1) should relieve immunosuppression (immune rheostat) activity in the tumor bed. Note that, for clarity, only the primary interactions of PD-1, PD-L1, and PD-L2 are illustrated. (C) Bispecific antibodies against CD3 passively recruit cytotoxic T cells to tumor cells. Blinatumomab is a single-chain bispecific antibody that is composed of an anti-CD3 arm that recognizes the T cell receptor and an anti-CD19 arm that recognizes a surface antigen on the surface of B cell lymphoma cells; diagrammed is a conventional bispecific IgG for clarity. Recruitment of the T cells to the tumor cells in this way results in efficient tumor cell killing, as if the T cell had recognized its cognate peptide-MHC on the tumor cell target.

Following closely behind the success of ipilimumab are mAbs that antagonize the interaction of programmed death–1 (PD-1), another negative regulator of T cells, with its ligands PD-L1 and PD-L2 (Fig. 3B) (48). Two antibodies against PD-1 (nivolumab and lambrolizumab) and one to PD-L1 (MPDL3280A) are the most advanced in the clinic. Although not as powerful of a brake as CTLA4, PD-1’s role is to prevent the unrestrained activation of T cells that have been previously activated. PD-L1 is often expressed by tumor cells or immune cells infiltrating into tumors. Evolutionarily, its role probably developed as a regulatory node in viral immunity, preventing T cell hyperactivation and bystander killing of noninfected cells during virus infection: For example, normal cells react to interferon-γ (IFN-γ) release by local effector T cells by expressing PD-L1 to protect themselves against T cell killing. In cancer, tumor cells are thought to react in much the same way upon entry of tumor-specific IFN-γ–secreting cytotoxic T cells. Thus, the binding of tumor PD-L1 to PD-1 on activated T cells inhibits their effector function. Mechanistically, this appears to occur by the recruitment of a phosphatase to the PD-1 cytoplasmic domain that inactivates the phosphatidylinositol 3-kinase pathway in T cells, exhausting their ability to generate or release cytotoxic granules (49). Thus, the PD-1/PD-L1 axis serves as a front-line mechanism of immune suppression in the tumor bed. PD-L2 in particular is also expressed by dendritic cells and may play a role in maintaining peripheral tolerance.

Both preclinically and clinically, the addition of mAbs to either PD-1 or PD-L1 blocks their interaction, thereby rescuing T cell cytotoxic activity; often, this results in rapid and substantial tumor shrinkage coupled with long-term, durable responses (48). Presumably because the primary mode of action is to overcome immunosuppression of antitumor T cells in the tumor bed (as opposed to anti-CTLA4, which amplifies the production of T cells of all specificities in lymph nodes), adverse events associated with PD-1/PD-L1 blockade appear substantially less serious than with ipilimumab. Although registrational trials are just now getting under way, preliminary suggestions are that mAbs against both PD-1 and PD-L1 are active (5054); it is too soon to know whether one or the other approach will distinguish itself on the basis of safety or efficacy. In this respect, it is interesting to note that PD-1 blockade will also inhibit interactions of T cells with PD-L2 on antigen-presenting cells (especially in the lung), which may or may not increase chances for toxicity; pneumonitis is a risk factor in patients treated with nivolumab, anantibody against PD-1. In addition, the PD-1 antibodies in the clinic are of the human IgG4 subclass with reduced potential for ADCC compared with the PD-L1 antibody, which has been engineered to completely eliminate FcγR binding and has yet to elicit serious pneumonitis (53). This difference or the difference in target (PD-1 versus PD-L1) may or may not correlate with increased efficacy or safety.

Given the complementary mechanisms of action of ipilimumab and the anti–PD-1/L1 antagonists, a recent clinical study has demonstrated that the two agents can show impressive additive activity, as well as additive toxicity, in melanoma (55). Because different criteria were used to judge clinical responses in this versus single-antibody trials [e.g., use of Response Evaluation Criteria in Solid Tumors (RECIST) versus World Health Organization (WHO) criteria for tumor shrinkage], further study will be required to understand the degree of enhanced benefit.

Antibodies that block immune checkpoints and immune suppression in the tumor bed have produced long-term, durable patient responses rarely seen with other therapeutics and, as such, may again change the face of cancer therapy similar to the initial antibodies against EGFR and HER2 15 years ago. It is likely that the coming years will see many more antibodies in this space as additional regulatory targets, both cell-associated and secreted, are identified and investigated.

Other Related Approaches

At some level, all therapies involving mAbs are immune therapies in that the immune system was required to produce and engage the therapeutic. However, it is useful to regard immunotherapies as ones that seek to actively manipulate the immune system, most often the T cell compartment, as is the case for antibodies to CTLA4, PD-1, and PD-L1. Hybrid approaches are already being introduced. One example of note is a specific single-chain antibody, blinatumomab, comprising tandem single-chain Fv fragments that bind CD19 on lymphoma and normal B cells and CD3, the antigen receptor of T cells (56). Treatment of patients with this small protein is effective at depleting non–Hodgkin’s lymphoma cells by recruiting T cells to tumor cells and activating their cytotoxic effector function regardless of the T cell’s own inherent specificity. Although active in early clinical trials, it is rapidly cleared from the circulation because of its small size, necessitating continuous infusion. Future work will no doubt investigate modifying the platform to enable more convenient dosing (Fig. 3C).

Another approach accomplishes the same task, except by fusing the anti-CD3 specificity to soluble, recombinant T cell receptors that can be selected to detect tumor-specific peptide–major histocompatibility complex (MHC) class I molecules (57). As above, T cells are recruited regardless of their inherent specificity and triggered to become antitumor effectors.

A third approach involves genetically modifying a patient’s own T cells to express a membrane-bound antibody against tumors fused to signaling molecules that trigger T cell killing when the modified T cell detects a cognate tumor cell. The T cells (CARs) have yielded a promising result, at least in hematologic cancer in early clinical trials (58).

Last, conventional antibodies can be generated with engineered Fc domains that increase, rather than decrease, FcγR binding, in order to recruit macrophages and NK cells to mediate ADCC (59).

The Past, Present, and Future

Antibody therapeutics in cancer have a rich history, an exciting present, and a promising future. Although it is uncertain what new platforms will emerge as being efficacious and useful, it is certain that many new approaches will be tried in the years to come. The use of antibody therapeutics in combination with each other is also emerging. Indeed, compelling phase I data have recently been reported for dual immune checkpoint therapy using ipilimumab and nivolumab (55). These combinations have the potential to significantly lower or, in some cases, eliminate the amount of cytotoxic chemotherapy that is still currently the backbone of most oncology treatments. Systemic treatment of patients, which includes targeted mAb therapy, before surgery is also emerging (32). Moreover, one could imagine scenarios where these approaches are likely to reduce the need for extensive surgery. New delivery platforms, new approaches to ADCs, and new ideas to manipulate the immune response or the tumor microenvironment are certainly on the horizon and bode well for a renaissance of interest in antibody biochemistry and function, as well as for cancer patients.

References and Notes

  1. Acknowledgments: M.S. and I.M. are full-time employees of Genentech, Incorporated, a member of the Roche Group. Both authors have been granted options and own shares of Roche Holding AG. Erlotinib, rituximab, trastuzumab, bevacizumab, pertuzumab, and ado-trastuzumab emtansine are marketed products from Genentech, Incorporated. MEHD7945A, onartuzumab, and MPDL3280A are under clinical development by Genentech, Incorporated and are not FDA approved.
View Abstract

Stay Connected to Science

Navigate This Article